how
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies and processes untrusted external source code and metadata to generate architectural walkthroughs, creating a potential surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the context when the skill reads "unfamiliar code" and when "explorer findings" are interpolated into the explainer template in
references/explainer-prompt.md. - Boundary markers: The templates lack explicit boundary markers or instructions to the agent to disregard instructions embedded within the analyzed code comments or string literals.
- Capability inventory: The skill uses file-read operations and dispatches the
oh-my-claudecode:exploretool to map codebase structures. - Sanitization: There is no evidence of sanitization or escaping of the code content before it is processed by the language model roles.
Audit Metadata